The British Journal of Radiology, 85 (2012), 897–904

Multidetector CT features of pulmonary focal ground-glass

opacity: differences between benign and malignant
L FAN, MD, S-Y LIU, MD, Q-C LI, MD, H YU, MD and X-S XIAO, MD

Department of Radiology, Changzheng Hospital, Second Military Medical University, Shanghai, China

Objective: To evaluate different features between benign and malignant pulmonary

focal ground-glass opacity (fGGO) on multidetector CT (MDCT).
Methods: 82 pathologically or clinically confirmed fGGOs were retrospectively
analysed with regard to demographic data, lesion size and location, attenuation value
and MDCT features including shape, margin, interface, internal characteristics and
adjacent structure. Differences between benign and malignant fGGOs were analysed
using a x2 test, Fisher’s exact test or Mann–Whitney U-test. Morphological
characteristics were analysed by binary logistic regression analysis to estimate the
likelihood of malignancy.
Results: There were 21 benign and 61 malignant lesions. No statistical differences
were found between benign and malignant fGGOs in terms of demographic data, size,
location and attenuation value. The frequency of lobulation (p50.000), spiculation
(p50.008), spine-like process (p50.004), well-defined but coarse interface (p50.000),
bronchus cut-off (p50.003), other air-containing space (p50.000), pleural indentation
(p50.000) and vascular convergence (p50.006) was significantly higher in malignant
fGGOs than that in benign fGGOs. Binary logistic regression analysis showed that
lobulation, interface and pleural indentation were important indicators for malignant
diagnosis of fGGO, with the corresponding odds ratios of 8.122, 3.139 and 9.076, Received 22 August 2010
respectively. In addition, a well-defined but coarse interface was the most important Revised 14 March 2011
indicator of malignancy among all interface types. With all three important indicators Accepted 17 March 2011
considered, the diagnostic sensitivity, specificity and accuracy were 93.4%, 66.7% and
DOI: 10.1259/bjr/33150223
86.6%, respectively.
Conclusion: An fGGO with lobulation, a well-defined but coarse interface and pleural ’ 2012 The British Institute of
indentation gives a greater than average likelihood of being malignant. Radiology

With the availability of low-dose spiral CT scan of the
lung, focal ground-glass opacity (fGGO) that was difficult
to detect on conventional chest radiographs has increas-
ingly been detected [1–3]. Ground-glass opacity (GGO) is
defined as an area of a slight homogeneous increase in
density, which does not obscure underlying bronchial
structures or vascular margins on high-resolution CT
(HRCT) [4]. Pathologically, GGO may be caused by partial
airspace filling, interstitial thickening with inflamma-
tion, oedema, fibrosis, neoplastic proliferation, the normal
respiratory condition or increased pulmonary capillary
blood volume [5]. GGO can be classified as pure GGO
(pGGO) or mixed GGO (mGGO) based on the presence
of solid components. Although GGO is a common and
Address correspondence to: Dr Shi-yuan Liu, Department of
Radiology, Changzheng Hospital, Second Military Medical Univer-
sity, No. 415 Fengyang Road, Shanghai 200003, China. E-mail:
lsy0930@163.com
This work was funded by the Natural Science Foundation of
Shanghai (no. 10ZR1438900), the Youth Fund of the National
Natural Science Foundation of China (no. 81000602), the
‘‘Mountaineering plan’’ Major Subject Program of the Science and
Technology Committee of Shanghai (no. 06DZ19503), the Major
Subject Program of the Science and Technology Committee of
Shanghai (no. 08DZ1900707) and National the Natural Science
Foundation of China (no. 30970800).
non-specific finding of lung HRCT, and may occur in
benign lung conditions such as organising pneumonia,
focal fibrosis and haemorrhage [6–8], it has recently
received considerable attention because it may indicate
an early underlying lung cancer, which in most cases
presents as bronchioloalveolar carcinoma (BAC) and
adenocarcinoma with a predominant BAC component. It
was reported in a study [9] that 17 of 28 pGGOs were BAC,
3 were adenocarcinoma and 8 were atypical adenomatous
hyperplasia (AAH). Several other studies [10, 11] have also
indicated that mGGOs are more likely to be malignant,
with the malignant rate of pGGO and mGGO being 18%
and 63%, respectively [12]. The aim of the present study
was to retrospectively compare the features of benign and
malignant fGGOs on thin section multidetector CT
(MDCT) images in an attempt to identify characteristics
that would help the differential diagnosis of fGGOs.
Methods and materials
Patient population
From January 2007 to April 2010, a consecutive series of
84 fGGO lesions of #4 cm in diameter were found on CT

The British Journal of Radiology, July 2012 897


in 84 patients in our institution. Of them, 82 fGGOs
in 82 patients that had been confirmed by operative
pathology (n557), biopsy pathology (n512) or clinical
diagnosis (n513) by June 2010 were used for retrospective
analysis in this study. The clinical diagnosis was based on
a significant reduction in size or complete disappea-
rance of the lesion on follow-up MDCT images after
anti-inflammatory therapy. The other two lesions were
excluded from the present study owing to their equivocal
nature (i.e. that they have remained stable during the
follow-up period until now). Of the 82 fGGOs enrolled, 61
were diagnosed as malignant, comprising adenocarci-
noma (26), BAC (33), large-cell lung cancer (1) and
adenosquamous carcinoma (1); and the remaining 21
were diagnosed as benign, comprising inflammation (14),
organising pneumonia (1), focal fibrosis (2) and AAH (4).
Definite histological or clinical diagnosis of these 82
fGGOs was obtained within 6 months. The local ethics
committee approved this retrospective study and waived
informed consent.
Image acquisition
All the patients were scanned on a 16-MDCT scan-
ner (Acquilion 16; Toshiba, Tokyo, Japan). Breath-hold
training was carried out before each examination. All
subjects were asked to hold their breath at the end of
inspiration for as long as possible. A non-enhanced scan
was performed from the thoracic inlet to the middle
portion of the kidneys, followed by contrast-enhanced
scan of the lesion area only in arterial and delayed
phases, to reduce exposure to radiation. For contrast-
enhanced scans, all injections were performed with an
automatic power injector, with which 90 ml contrast
medium (Optiray 350 mgI ml–1; Mallinckrodt Medical,
St Louis, MO) was injected into the antecubital vein
at a rate of 4 ml s–1. Arterial phase and delayed phase
images were acquired at 20–25 s and 90 s after injection,
respectively [13]. The imaging parameters were as
follows: 250 mA, 120 kVp, collimation 1 mm, pitch 0.98
and filtration coefficient 52 or 1 (high- or standard-
resolution algorithms).
Demographic data and image analysis
Patients’ demographic data included sex and age. All
the raw data sets were sent to a Vitrea 1 workstation and
reconstructed. Reconstruction for axial images was per-
formed with a section thickness of 1 mm, a reconstruction
interval of 0.8 mm, field of view (FOV) of approximately
18618 cm to 20620 cm and a filtration coefficient of 52 or 1,
and was viewed at the mediastinal and pulmonary
window settings. Multiplanar reconstructions (MPRs)
and three-dimensional volume-rendered (VR) images
were then performed to display the morphological
features and relationships with the adjacent bronchi and
vessels.
All the post-processed images were interpreted by two
thoracic radiologists with 15 and 10 years’ experience in
chest CT. The observers were blinded to the subjects’
identities and clinical data. Decisions on CT findings
were reached by consensus. CT scans were assessed by
898
L Fan, S-Y Liu, Q-C Li et al
observing lesion location, size, shape, margin, interface,
internal characteristics, adjacent structure and attenua-
tion value. Lesion size was defined as the longest dimen-
sion. Shape was classified as irregular or round/oval.
Marginal characteristics included lobulation, spiculation,
cusp angle and spine-like process. Cusp angle and spine-
like process were both defined as a structure extend-
ing from the lesion, while their borders with lung
parenchyma were different: cusp angle with either
straight or slightly concave borders, and spine-like pro-
cess with at least one convex border. Interfaces were
classified as one of three types: ill-defined, well-defined
and smooth, and well-defined but coarse. Internal
characteristics included air bronchograms and other
air-containing space. Air bronchograms were classified
as natural, dilated and distorted, or cut-off. The defini-
tion of a natural air bronchogram was running naturally
with the regular bronchial wall [14]. Findings of adjacent
structures included the pleural indentation sign and the
vascular convergence sign. For mGGO, attenuation
values of the GGO component, solid component and
adjacent parenchyma on non-enhanced scan were
recorded, and those of the solid component on con-
trast-enhanced scan were also recorded. For pGGO,
attenuation values of the GGO and adjacent parenchyma
were recorded only on non-enhanced scan. Then, the
difference in attenuation value between the GGO
component and adjacent parenchyma was calculated. A
region of interest (oval or circular) covering one-half to
two-thirds of the largest area in a lesion away from air-
containing space was selected.
Statistical analysis
The statistical analysis of all data sets was performed
with SPSS v.11.0 software (SPSS Inc., Chicago, IL). The
sex ratio of the patients and the morphological features
of the lesions were analysed for differences between
benign and malignant fGGOs using the x2 test or Fisher’s
exact test. Patient age, lesion size and attenuation value
were analysed with the Mann–Whitney U-test. A two-
sided value of p,0.05 was used as the criterion to
indicate a statistically significant difference.
In addition, binary logistic regression analysis was
performed on the morphological characteristics to draw a
regression equation to estimate the likelihood of malig-
nancy. The forward conditional method was employed
for variable selection.
Results
Demographic data, lesion size and location
With respect to the demographic data of the 82 patients
(35 males, 47 females; mean age, 59.02¡9.63 years;
age range, 40–84 years), there was no significant difference
in either the sex ratio (p50.120) or age (p50.648) between
patients with benign and malignant fGGOs. Also, there
was no significant difference in either lesion size (mean
size, 2.35¡0.72 cm; size range, 0.5–4 cm, p50.955) or
location (p50.902) between benign and malignant fGGOs
(Table 1).
The British Journal of Radiology, July 2012
Differential diagnosis of pulmonary focal ground-glass opacity by MDCT

Table 1. Demographic data, nodule size and location of space (14.3 vs 59.0, p50.000; Figure 5), pleural indentation
benign and malignant focal ground-glass opacity (fGGO) (4.8 vs 70.5, p50.000; Figure 1) and vascular convergence
(4.8 vs 36.1, p50.006; Figure 1) (Table 3). In terms of
Benign Malignant
fGGO (n521) (n561) p-value
air bronchograms, there was no significant difference
between benign and malignant fGGOs in the incidence of
Sex ratio (M:F) 12:9 23:38 0.120a either natural (p50.502) or dilated and distorted (p51.000)
Average age (years) 60.5¡11.07 58.68¡9.39 0.648b bronchus, while the incidence of cut-off was significantly
Average size D (cm) 2.33¡0.71 2.36¡0.73 0.955b
different (0 vs 32.8, p50.003; Figure 1). There was no
Location
RUL 10 21 0.902c significant difference between benign and malignant
RML 2 7 fGGOs in the incidence of other morphological character-
RLL 3 10 istics such as shape, well-defined and smooth interface
LUL 4 13 and cusp angle (Figure 6) (p.0.05).
LLL 2 10
D, longest dimension; F, female; LLL, left lower lobe; LUL, left
upper lobe; M, male; RLL, right lower lobe; RML, right Binary logistic regression analysis
middle lobe; RUL, right upper lobe.
Sex ratio and location are expressed in terms of frequency; In binary logistic regression analysis of the present
age and size are expressed as mean ¡ standard deviation. study, benignity and malignancy were regarded as
a 2
x test. dependent variables, and the shape, lobulation, spicula-
b
Mann–Whitney U-test. tion, cusp angle, spine-like process, interface, air bronch-
c
Fisher’s exact test. ograms, other air-containing space, pleural indentation
sign and vascular convergence were regarded as
Attenuation values
independent variables. The result showed that lobula-
The fGGOs included 68 lesions with mGGO appear- tion, interface and pleural indentation were important
ance and 14 with pGGO appearance. For the solid indicators for malignant diagnosis of fGGO; the cor-
component of the mGGOs, the mean CT attenuation fresponding regression equation was as follows:
values on non-enhanced CT images (p50.520), on arterial
phase images (p50.464) and on delayed phase images lnðp=1{pÞ~{2:268z2:095|lobulationz1:144|
(p50.869) were not significantly different between
benign and malignant mGGOs. For the difference in interfacez2:206|pleural indentation
attenuation value between the GGO component and
the adjacent parenchyma on non-enhanced CT images, where p is the probability of malignant fGGO. When the
there was no significant difference between benign and p-value was $0.5, the lesion was expected to be
malignant mGGOs or pGGOs, the corresponding p- malignant, while the others were categorised as benign.
values being 0.068 and 1.000, respectively (Table 2). In terms of malignant fGGO diagnosis, the odds ratios
(ORs) of lobulation, pleural indentation and interface
Comparison of morphological features between were 8.122, 9.076 and 3.139, respectively (Table 4). In
benign and malignant fGGOs other words, the malignant risk of an fGGO with
lobulation was 8.122 times that without lobulation; the
Of the 21 benign fGGOs, 14 were mGGOs and 7 were malignant risk of an fGGO with pleural indentation was
pGGOs. Of the 61 malignant fGGOs, 54 were mGGOs 9.076 times that without pleural indentation; the malig-
and 7 were pGGOs. Statistically significant difference was nant risk of an fGGO with well-defined but coarse
found between benign and malignant fGGOs in the interface was 3.139 times that with well-defined and
frequency of lobulation (14.3 vs 83.6, p50.000; Figure 1), smooth interface; and the malignant risk of an fGGO
spiculation (4.8 vs 34.4, p50.008; Figure 2), spine-like with well-defined and smooth interface was 3.139 times
process (0 vs 29.5, p50.004; Figure 3), ill-defined interface that with ill-defined interface. It was evident that well-
(66.7 vs 1.6, p50.000; Figure 4), well-defined but coarse defined but coarse interface was the most important
interface (33.3 vs 93.4, p50.000; Figure 2), air-containing discriminator of the three interface types.

Table 2. Attenuation values of benign and malignant mixed ground-glass opacity (mGGO) and pure ground-glass opacity (pGGO)
Attenuation value (HU)

fGGO Component Benign Malignant p-value

mGGO Solid component
Non-enhanced scan 37.57¡9.34 38.75¡7.79 0.520b
Arterial phase 65.14¡21.71 60.85¡24.90 0.464b
Delayed phase 80.00¡23.09 79.55¡24.91 0.869b
GGO componenta 482.57¡144.13 341.93¡198.26 0.068b
pGGO GGO componenta 210¡54.06 314.5¡62.55 1.000b
fGGO, focal ground-glass opacity.
Attenuation values are expressed as mean ¡ standard deviation.
a
Difference of attenuation value between the GGO component and adjacent parenchyma.
b
Mann–Whitney U-test.

The British Journal of Radiology, July 2012 899

 L Fan, S-Y Liu, Q-C Li et al

(a) (b)

Figure 1. Bronchioloalveolar cardnoma in a 62-year-old male. (a) Oblique multiplanar reconstruction lung window shows a
23 mm focal ground-glass opacity with lobulation and bronchus cut-off in the right upper lobe. (b) Maximum-intensity
projection shows the vascular convergence sign and pleural indentation.

The probability of a malignancy was calculated for each
lesion using the above-mentioned regression equation.
Table 5 lists the numbers of expected diagnosis by the
logistic regression equation (expected) and final patholo-
gical or clinical diagnosis by follow-up (observed) of
benign and malignant fGGOs. Predictive values of
the regression analysis were 14 and 57 for benign and
malignant fGGO, respectively. Sensitivity, specificity and
accuracy with this regression equation were calculated as
93.4%, 66.7% and 86.6%, respectively.
Binary logistic regression analysis of the morphological
characteristics was also performed between benign and
malignant mGGOs, and between benign and malignant
pGGOs.
The results showed that lobulation and interface were
important indicators for malignant diagnosis of mGGO; the

(a) (b)

Figure 2. Adenocarcinoma in a 54-year-old female. (a) Transverse lung window thin-section (1 mm thick) and (b) coronal
multiplanar reconstruction show a 25 mm focal ground-glass opacity with spiculation and a well-defined but coarse interface in
the right lower lobe.

900 The British Journal of Radiology, July 2012

Differential diagnosis of pulmonary focal ground-glass opacity by MDCT
Figure 3. Bronchioloalveolar carcinoma in a 45-year-old
female. Sagittal multiplanar reconstruction lung window
shows a 28 mm focal ground-glass opacity with a spine-like
process (arrow) in the right upper lobe.
regression equation was:
lnðp=1{pÞ~{2:884z2:977|lobulationz
1:673|interface
In terms of malignant mGGO diagnosis, the ORs of
lobulation and interface were 19.620 and 5.329, respectively.
(a)
Figure 4. An inflammatory lesion in a 53-year-old female. (a) Transverse lung window shows a focal ground-glass opacity with
an ill-defined interface in the right lower lobe. (b) Transverse lung window shows that the lesion has almost disappeared after
anti-inflammatory therapy.
The British Journal of Radiology, July 2012
Sensitivity, specificity and accuracy were 98.1%, 57.1% and
89.7%, respectively.
The results of binary logistic regression analysis also
showed that interface was the most important indicator
for malignant diagnosis of pGGO; the regression equation
was:
lnðp=1{pÞ~{1:657z1:522|interface
The OR was 4.582 for the malignant pGGO diagnosis.
Sensitivity, specificity and accuracy were calculated as
57.1%, 71.4% and 64.3%, respectively.
Discussion
Ever since spiral CT was used in lung cancer screen-
ing, small peripheral or early-stage lung cancers have
been detected more frequently. Histologically, most of
these lung cancers are well-differentiated adenocarci-
noma or BAC [1, 2, 6, 15], typically appearing as fGGOs
on HRCT. About 30% of benign pulmonary lesions also
manifest as fGGOs [16]. Although some studies have
reported MDCT features of fGGO lesions [8, 17–19], the
differential features between benign and malignant
fGGOs still deserve to be investigated.
The binary regression analysis of our study revealed
that lobulation, interface and pleural indentation were
the most important discriminators between benign and
malignant fGGOs. The corresponding regression equa-
tion for differential diagnosis of fGGOs yielded high
sensitivity (93.4%) and accuracy (86.6%).
The marginal characteristics of pulmonary nodules
usually reflect their underlying pathological nature.
Lobulation is a common finding of lung MDCT and
is more frequently seen in malignant than in benign
(b)
901
 L Fan, S-Y Liu, Q-C Li et al

to that reported in another study [21], while the frequency

Figure 5. Adenocarcinoma in a 59-year-old male. Transverse
lung window thin-section (0.5 mm thick) shows an 8 mm
focal ground-glass opacity with air-containing space in the
right middle lobe.
pulmonary lesions. Pathologically, malignant lobulation
is caused by different growth velocity owing to different
cell differentiation, tumour growth blocked by the adjacent
pulmonary interstitium and contraction of fibrous tissue
inside the lesion [5, 18, 20]. The pathological basis of
benign lobulation is hyperplasia of connective tissue
inside or around the nodules and cicatrical contraction
[8, 20]. It was found in the present study that the fre-
quency of lobulation in malignant fGGOs was signifi-
cantly higher than in benign ones (83.6% vs 14.3%). The
frequency of lobulation in malignant fGGOs was similar
of lobulation in benign fGGOs was lower. The reason may
be due to the small sample size of benign fGGOs in this
study.
Interface is another valuable discriminator for benign
and malignant fGGOs. In the present study, we classified
the interface as ill-defined, well-defined and smooth,
and well-defined but coarse. The result of our binary
regression analysis showed that a well-defined but
coarse interface was the most important discriminator
among the three types. In most cases of lung cancer, the
interface is usually well defined but wholly or partially
coarse, the pathological basis of which generally includes
the following three aspects: an infiltrative tumour
growth, inflammatory reaction in the peritumour par-
enchyma, and carcinomatous embolus formation in
small vessels or lymph vessels [8, 18]. Of the three
aspects, the tumour growth pattern is the most impor-
tant. In the present study, except for one case with an ill-
defined interface, all malignant cases appeared with a
well-defined interface (57 coarse, 3 smooth). Conversely,
the interface of benign lesions is usually ill defined
owing to infiltration of inflammatory cells. Two-thirds
(14/21) of the benign fGGOs in our series showed an ill-
defined interface. The frequency of well-defined and
smooth interface in benign and malignant fGGOs was
very similar (p50.566). It is pathologically caused by a
stacking tumour growth or a pseudocapsule resulting
from compression of the adjacent pulmonary parench-
yma by fast tumour growth [19]. In our opinion, a
well-defined but coarse interface is more suggestive of
malignancy, while an ill-defined interface is more likely
to be benign. Nambu et al [18] pointed out that a well-
defined margin was a valuable discriminator for GGO.

Table 3. Distribution of morphological features in benign and malignant focal ground-glass opacity on mutidetector CT
Benign (n521) Malignant (n561)

fGGO mGGO pGGO Total 1 mGGO pGGO Total 2 p-value

Shape
Irregularity 10 1 11 (52.4) 13 0 13 (21.3) 0.07a
Round/oval 4 6 10 (47.6) 41 7 48 (78.7)
Margin
Lobulation 2 1 3 (14.3) 48 3 51 (83.6) 0.000a
Spiculation 1 0 1 (4.8) 21 0 21 (34.4) 0.008a
Cusp angle 2 0 2 (9.5) 2 0 2 (3.3) 0.568b
Spine-like process 0 0 0 (0) 18 0 18 (29.5) 0.004b
Interface
Ill defined 9 5 14 (66.7) 1 0 1 (1.6) 0.000b
Well defined, smooth 0 0 0 (0) 0 3 3 (4.9) 0.566b
Well defined, coarse 5 2 7 (33.3) 53 4 57 (93.4) 0.000b
Internal characteristics
Air bronchograms
Natural 3 2 5 (23.8) 9 0 9 (14.8) 0.502b
Dilated and distorted 3 0 3 (14.3) 9 0 9 (14.8) 1.000b
Cut-off 0 0 0 (0) 19 1 20 (32.8) 0.003a
Air-containing space 3 0 3 (14.3) 33 3 36 (59.0) 0.000a
Adjacent structure
Pleural indentation 1 0 1 (4.8) 40 3 43 (70.5) 0.000a
Vascular convergence 1 0 1 (4.8) 22 0 22 (36.1) 0.006a
fGGO, focal ground-glass opacity; mGGO, mixed ground-glass opacity; pGGO, pure ground-glass opacity.
Data are expressed in terms of frequency or percentages (in parentheses).
a 2
x test.
b
Fisher’s exact test.

902 The British Journal of Radiology, July 2012

Differential diagnosis of pulmonary focal ground-glass opacity by MDCT

(a) (b)

Figure 6. An inflammatory lesion in a 75-year-old male. (a) Transverse lung window thin-section (1 mm thick) shows a 25 mm
focal ground-glass opacity in the right middle lobe. (b) Sagittal multiplanar reconstruction shows cusp angle (white arrow).

Although interface is an important discriminator for
differential diagnosis, focal fibrosis, organising pneumo-
nia and mucin-producing tumour should be excluded.
Focal fibrosis and organising pneumonia with a well-
defined interface, which are caused by the alveolar
septum thickening due to fibrosis and the air-containing
space is similar to the MDCT manifestation of adeno-
carcinoma [22–24]. An ill-defined interface may occur in
peripheral lung cancer, with an incidence of about 10%,
which is an atypical presentation and mainly occurs in
mucin-producing BAC. Mucin accumulation in the
alveolar space around the tumour results in an ill-
defined interface, which resembles inflammation.
The pleural indentation sign was also a common and
valuable indicator for differential diagnosis of malig-
nancy. The prerequisites of pleural indentation include
the following two aspects: no conglutination between the
parietal pleura and visceral pleura, and a distance
between of lesion and pleura .2 cm. Based on these
prerequisites, the contraction of fibrous tissue inside the
lesion leads to pleural indentation [20, 25]. The frequency
of pleural indentation was significantly higher in malig-
nant fGGOs than in benign fGGOs (70.5% vs 4.8%). In
addition, it mainly occurred in mGGOs. Nambu et al [18]
hypothesised that it may be attributed to the strong
contraction of the solid component.
The frequency of spiculation, spine-like process, bron-
chus cut-off, other air-containing space and vascular
convergence in malignant fGGOs was significantly higher
than in benign fGGOs, though they are not the important
risk factors for malignancy. Moreover, no spine-like process
or bronchus cut-off was observed in benign fGGOs. It is
therefore assumed that either spine-like process or
bronchus cut-off might be specific for malignancy. With
respect to the demographic data, lesion size, location and
attenuation value, no differences were found between
benign and malignant fGGOs, which is consistent with
previous studies [11, 19, 26].
There are several limitations in the present study, which
suggests that further investigation is warranted. First of all,
with regard to case selection, the sample size of benign
fGGOs was small, especially for pGGO; the follow-up
duration was relatively short; and most of the benign
fGGOs were diagnosed according to size reduction or
complete disappearance of the lesion as evidenced by the
anti-inflammatory therapy. However, some benign and
many malignant fGGOs with long volume-doubling time
may remain unchanged in size and attenuation value even
over a long follow-up period. In the present study, two
lesions manifested stability during the follow-up period
and are still being followed up, which were excluded
from our study owing to their equivocal nature. Therefore,

Table 4. Binary logistic regression analysis of focal ground-glass opacity

Variables in the equation

Exp(B) 95% CI
Independent variables
and constant B SE Wals df Sig Exp(B) Lower Upper

Lobulation 2.095 0.854 6.018 1 0.014 8.122 1.524 43.301

Interface 1.144 0.490 5.448 1 0.020 3.139 1.201 8.200
Pleural indentation 2.206 1.180 3.492 1 0.062 9.076 0.898 91.736
Constant 22.268 0.790 8.246 1 0.004 0.103
B, regression coefficient; CI, confidence interval; df, degrees of freedom; Exp(B), eB5odds ratio; SE, standard error; Sig,
significance; Wals, x2 value.

The British Journal of Radiology, July 2012 903


more cases and longer follow-up period are needed in
future studies to obtain more accurate results. Also, the
reconstruction parameters and window settings were the
same for all fGGOs, which may result in possible missing
features, especially for small pGGOs. So different recon-
struction parameters and multiple middle window settings
should be applied in future studies to take the size and
attenuation value of fGGOs into consideration.
In conclusion, an fGGO nodule with lobulation, a well-
defined but coarse interface and pleural indentation gives
a greater than average likelihood of being malignant.
Acknowledgment
The authors would like to thank Professor Wei-hua Dong
(Second Military Medical University) for critical reading
and polishing the English language of the manuscript.
Reference
1. Sone S, Takashima S, Li F, Yang Z, Honda T, Maruyama Y,
et al. Mass screening for lung cancer with mobile spiral
computed tomography scanner. Lancet 1998;351:1242–5.
2. Henschke CI, McCauley DI, Yankelevitz DF, Naidich DP,
McGuinness G, Miettinen OS, et al. Early Lung Cancer
Action Project: overall design and findings from baseline
screening. Lancet 1999;354:99–105.
3. Tsubamoto M, Kuriyama K, Kido S, Arisawa J, Kohno N,
Johkoh T, et al. Detection of lung cancer on chest radio-
graphs: analysis on the basis of size and extent of ground-
glass opacity at thin-section CT. Radiology 2002;224:139–44.
4. Austin JH, Müller NL, Friedman PJ, Hansell DM, Naidich
DP, Remy-Jardin M, et al. Glossary of terms for CT of the
lung: recommendations of the Nomenclature Committee of
the Fleischner Society. Radiology 1996;200:327–31.
5. Nakajima R, Yokose T, Kakinuma R, Nagai K, Nishiwaki Y,
Ochiai A. Localized pure ground-glass opacity on high
resolution CT: histologic characteristics. J Comput Assist
Tomogr 2002;26:323–9.
6. Park CM, Goo JM, Lee HJ, Lee CH, Chun EJ, Im JG.
Nodular ground-glass opacity at thin-section CT: histologic
correlation and evaluation of change at follow-up.
Radiographics 2007;27:391–408.
7. Hara M, Oda K, Ogino H, Oshima H, Sato Y, Kiriyama M,
et al. Focal fibrosis as a cause of localized ground glass attenuation
(GGA): CT and MR findings. Radiat Med 2002;20:93–5.
8. Kim HY, Shim YM, Lee KS, Han J, Yi CA, Kim YK. Persistent
pulmonary nodular ground-glass opacity at thin-section CT:
histopathologic comparisons. Radiology 2007;245:267–75.
9. Nakata M, Saeki H, Takata I, Segawa Y, Mogami H, Mandai
K, et al. Focal ground-glass opacity detected by low-dose
helical CT. Chest 2002;121:1464–7.
10. Aoki T, Nakata H, Watanabe H, Nakamura K, Kasai T,
Hashimoto H, et al. Evolution of peripheral lung adeno-
carcinomas: CT findings correlated with histology and
tumor doubling time. AJR Am J Roentgenol 2000;174:763–8.
904
L Fan, S-Y Liu, Q-C Li et al
11. Li F, Sone S, Abe H, MacMahon H, Doi K. Malignant versus
benign nodules at CT screening for lung cancer: comparison
of thin-section CT findings. Radiology 2004;233:793–8.
12. Henschke CI, Yankelevitz DF, Mirtcheva R, McGuinness G,
McCauley D, Miettinen OS, et al. CT screening for lung
cancer: frequency and significance of part-solid and
nonsolid nodules. AJR Am J Roentgenol 2002;178:1053–7.
13. Yu H, Liu SY, Li HM, Xiao XS, Dong WH. Empirical
description of bronchial and nonbronchial arteries with
MDCT. Eur J Radiol 2010;75:147–53.
14. Cui Y, Ma DQ, Liu WH. Value of multiplanar reconstruc-
tion in MSCT in demonstrating the relationship between
solitary pulmonary nodule and bronchus. Clin Imaging
2009;33:15–21.
15. Lee HJ, Goo JM, Lee CH, Yoo CG, Kim YT, Im JG. Nodular
ground-glass opacities on thin-section CT: size change
during follow-up and pathological results. Korean J
Radiol 2007;8:22–31.
16. Jang HJ, Lee KS, Kwon OJ, Rhee CH, Shim YM, Han J.
Bronchioloalveolar carcinoma: focal area of ground-glass
attenuation at thin-section CT as an early sign. Radiology
1996;199:485–8.
17. Kim TJ, Goo JM, Lee KW, Park CM, Lee HJ. Clinical,
pathological and thin-section CT features of persistent
multiple ground-glass opacity nodules: comparison with
solitary ground-glass opacity nodule. Lung Cancer 2009;
64:171–8.
18. Nambu A, Araki T, Taguchi Y, Ozawa K, Miyata K,
Miyazawa M, et al. Focal area of ground-glass opacity and
ground-glass opacity predominance on thin-section CT:
discrimination between neoplastic and non-neoplastic
lesions. Clin Radiol 2005;60:1006–17.
19. Oh JY, Kwon SY, Yoon HI, Lee SM, Yim JJ, Lee JH, et al.
Clinical significance of a solitary ground-glass opacity
(GGO) lesion of the lung detected by chest CT. Lung
Cancer 2007;55:67–73.
20. Winer-Muram HT. The solitary pulmonary nodule. Radio-
logy 2006;239:34–49.
21. Zerhouni EA, Stitik FP, Siegelman SS, Naidich DP, Sagel SS,
Proto AV, et al. CT of the pulmonary nodule: a cooperative
study. Radiology 1986;160:319–27.
22. Ujita M, Renzoni EA, Veeraraghavan S, Wells AU, Hansell
DM. Organizing pneumonia: perilobular pattern at thin-
section CT. Radiology 2004;232:757–61.
23. Yang PS, Lee KS, Han J, Kim EA, Kim TS, Choo IW. Focal
organizing pneumonia: CT and pathologic findings. J Korean
Med Sci 2001;16:573–8.
24. Park CM, Goo JM, Lee HJ, Lee CH, Chung DH, Chun EJ,
et al. Focal interstitial fibrosis manifesting as nodular
ground-glass opacity: thin-section CT findings. Eur Radiol
2007;17:2325–31.
25. Kuriyama K, Tateishi R, Doi O, Kodama K, Tatsuta M,
Matsuda M, et al. CT-pathologic correlation in small
peripheral lung cancers. AJR Am J Roentgenol 1987;149:
1139–43.
26. Ohtsuka T, Watanabe K, Kaji M, Naruke T, Suemasu K. A
clinicopathological study of resected pulmonary nodules
with focal pure ground-glass opacity. Eur J Cardiothorac
Surg 2006;30:160–3.
The British Journal of Radiology, July 2012